Tweak3D heeft een aantal Neon 250 (dat eindeloos gedelayde-maar-nu-toch-bijna-verkrijgbare PowerVRSG dingetje van NEC/Videologic) gepost:

A source close to Videologic sent me these genuine benchmarks for the upcoming PowerVR Neon250 board as part of a press release. Here's the info:

'After a long wait, PowerVR Series2 for the PC, in the guise of Neon250, will start shipping at the end of August.

The simple fact is that the Neon 250 is the fastest OpenGL gaming card on the planet. Running on the pre-production silicon (about 5% slower than production), and rather early drivers (all recent driver work has been happening on the production silicon), we're clocking the following numbers (in fps) in 16-bit on a PIII 500 with 128MB RAM:

...en de cijfertjes zie je hier.

In MultiMediaMarket melden ze dat Guillemot de grootste stuurtjes maker ter wereld is. Dit hebben ze te danken aan het feit dat ze ThrustMaster hebben overgenomen. ThrustMaster was en is met 45% marktleider ij het maken van pc stuurtjes.

"Guillemot distributeerde de afgelopen 5 jaar al ThrustMasterproducten in Frankrijk. De kwaliteiten die ThrustMaster levert, zullen een perfecte 'match' zijn voor Guillemot", aldus directeur Claude Guillemot.

Ander nieuws over Guillemot is dat ze voortaan het Ferrari logo op hun producten mogen zetten.. Dus mocht je deze maand weer de loterij niet hebben gewonnen maar toch Ferrari wil rijden, is Guillemot een goede oplossing.

Ik kwam bij Matrox Users wat info tegen over de overklokbaarheid van de G400:

Look at the highlighted improvement on fillrate after overclocking. It can be as high as 20%. Another interesting result is that in both cases, the fillrate improvement is almost the same regardless of the CPU used. However, Intel Pentium III seems to have a much better utilization of the tremendous fillrate the G400 offers. This shows that G400 is clearly CPU-bounded. With more powerful next generation CPUs coming out, such as AMD Athlon and Intel Mercs, we can expect G400 performance to scale accordingly. On the other hand, let's hope that Matrox will once again flex his muscles on driver development, offloading more processing chores from the host CPU to G400. That will definitely improve the low-res performance and make G400 sell like hot cake!

Benchmarks e.d. vind je op mgatools.matroxusers.com/.

Gamers Depot heeft een review gepost van de Creative Webcam III USB. De cam heeft een 640x480px resolutie en wordt geleverd met een leuke verzameling zachtgoed:

The WebCam Monitor is really, in my opinion, the main application that sets this product apart from the rest of the "wanna-be" WebCam out there. This killer program let's you setup the camera in what I like to call "spy" mode. For example, let's say you'd like to keep an eye on your office while your away, well then you can set the camera to take "snapshots" in whatever time interval you'd like. And then, you can either have the program upload that image to your email, or to a webpage! This could come in very handy for those times when you leave a pet home during the day, and you'd like to see why you're house is always a mess when you get home! You can also set it to take an image when it senses motion.

Upside.com heeft een interview gepost met Linus Torvalds, de man acher Linux. Hier heb je alvast een vraag:

Upside: It strikes me that there were many points when you could have gone commercial or decided that the Linux operating system should become commercial, and you obviously decided not to do that. Why?

Torvalds: It was a question of my interests. I was always interested in doing Linux for the technical side. It didn't start as a project to make money, and when it became clear that it could make money, that wasn't why I did it. That would have changed my motive.

I'm happy that people are making money on Linux because I think it gives more depth to Linux. It brings in new motivations and new factors that wouldn't have been there if it [weren't] for the commercial end of the business. But when I started Linux, [going commercial] wasn't an option.

Ik kwam bij Ace's Hardware nog wat info tegen over de huidige malaise bij Cyrix (overgenomen door VIA, Gobi en Mojavi CPU designs gedumpt):

Ok so Steve has updated his site where REAL Cyrix people bitch about NSM's ineptness in handling them. As for the Gobi project it seems to be in a rather bad condition: the 5 clock 256K on-die L2 did not turn out to work very well and the MHz is much lower than originally planned. The yield is also very low to the point that they can hardly sample it. Also, with the people in the project fired, it will take a miracle to debug Gobi in time. The MTRR is said to be messed up and thus leads to very disappointing performance. The last word is that the current revision does not work much better than the previous one. [break] En van Steve's site heb je hier een stukje: [/break] The real story about Via's choice to layoff people has to do with the fact that our R&D burn rate is about $10 million/month. Once this fact was unearthed, Via had to rethink its strategy and start immediately reducing costs. In addition, the motivation level of the M3 team prior to the layoffs was nonexistent. Anyone truly knowledgeable about the Gobi project would have observed its motivation level to be "lower" than in previous months, but team focus was continuing to improve through the months of June and July. Problems were being resolved and fixes were being put in place. Progress was being made in spite of all the distractions of the day.

>From the day the National announcement was made in May, I have heard two basic themes:

1) M3 team members whining and asking, "What can Via do for me?"

2) Gobi team members asking, "What can I do to make Cyrix and Via successful?"

The Gobi team has never received one offer of assistance from any M3 team members. That is, until several M3 team members were "asked" to join the Gobi project immediately after the mass layoff. This behavior is what caused the company to fall apart more than anything; lack of teamwork and focusing on things on which we had absolutely no influence. A sad state of affairs in a company which had many chances over the years to "eat Intel's lunch."

IBM heeft een PowerPC plankje ontwikkeld waarmee goedkoop een PPC powered Linux systeem gebouwd kan worden (tot op heden moest je daar een complete Mac voor aanschaffen). IBM gaat het moederbord niet zelf produceren, maar mobobakkers kunnen het design kosteloos overnemen:

IBM will not charge license fees or royalties, nor will it manufacture its own motherboards based on the design, Faure says. The company will provide schematics, component lists, and circuit board art to interested manufacturers, Faure adds. "It uses much of the technology in x86 motherboards, but with the PowerPC chip," Faure says. He would not say whom IBM is speaking with about the motherboards. Products based on the motherboards could become available early next year, he says.

The architecture harkens from IBM's efforts in the mid-90s to create reference designs for Macintosh clones, Faure says. IBM updated that work by adding things like higher bus speeds and support for accelerated graphics port signals. IBM showed a prototype of the motherboards at LinuxWorld Conference and Expo this week in San Jose, and will also show them next week at HP World in San Francisco. Currently, Apple is the only major computer manufacturer using PowerPC chips on its motherboards.

"A lot of people out there right now are running Linux on PowerPC but they're using Macs, so they're basically paying for a box to run both Mac software and Linux," Faure says.

If other manufacturers make PowerPC motherboards for Linux, people will have more choices for Linux platforms besides x86 PCs, Apple's PowerPC systems.

"It's going to reduce the cost of getting Linux on PowerPC," says Jason Haas, marketing director for LinuxPPC, which makes the primary distribution of Linux for PowerPC computers.

Apple didn't comment on the action by its PowerPC partner, but Faure says he'd "be surprised if they had any objection." Most Apple customers aren't Linux customers, he adds.